Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-08-03 · high
The same release fixes provider configs for reasoning models, broken since fields like reasoning_text arrived.
Why now: The fix landed August 1, addressing a failure mode that's easy to hit whenever agent sessions run unattended for hours.
Opencode 1.18.11, released August 1, fixes a bug where MCP connections over SSE could get stuck in an infinite reconnect loop. That happened after the MCP server sent back an error response. One misbehaving MCP server was enough to spin an agent session indefinitely, per the release notes.
That's a real trap for anyone running opencode as an agent harness with tool servers attached. No crash, no obvious error to grep for. Just a session that never finishes because one dependency keeps failing to reconnect cleanly.
The same release fixes provider model configs that use interleaved reasoning fields like reasoning_text, which had broken reasoning-model integrations. Two different bug classes closed in one point release: transport-layer connection handling and provider-layer config parsing.
Neither fix is dramatic. The release notes don't say how long the reconnect bug existed, or how many hung sessions it caused before someone caught it. But a client that hangs indefinitely instead of erroring out is exactly the failure you notice only after running an agent unattended for hours.
The more useful read here isn't about opencode specifically. Reconnect-after-error is a structural failure mode for any SSE-based MCP client. If you're running MCP servers through a different harness, check whether the same gap exists there too.
Each link below shares sources, entities, or timing with this story.
Raj Nagulapalle's FetchSandbox MCP took 107 votes on August 23, wiring 70+ API sandboxes into Cursor or Claude Code via MCP config. The claim is narrower and more testable than most agent tooling: reproduce the real integration failure against a sandbox, apply the fix, re-run...
There's been a fundamental gap in AI coding workflows that's been driving me crazy: your agent writes React components, generates CSS, builds entire UIs, but can't see the result. It's coding blind. The Chrome DevTools MCP server from Google's ChromeDevTools team fixes this, a...
EchoVault solves the "agent amnesia" problem — coding agents forgetting everything between sessions. It runs as an MCP server giving agents three tools: memory_context (load prior decisions), memory_search (find specific memories), and memory_save (persist learnings). Storage...
MCP server exposing 150+ cybersecurity tools for AI agents: nmap, sqlmap, ffuf, and 147 others. Covers pentesting, vulnerability discovery, and bug bounty automation. Works with Claude, GPT, Copilot — any MCP-compatible agent. Docker support included. Add it to your MCP config...
The 1.4.0a2 alpha ships langchain.mcp, an MCPAdapter turning any MCP server into tools you hand straight to create_agent. Connection handling is FastMCP's rather than a reimplementation, so a URL, a local stdio script, an in-process FastMCP server, a multi-server config, or a...
Waishnav/devspace (4,247 stars, v1.0.8 August 25) runs a local Node server exposing read, edit, search and shell execution over MCP, reached through a reverse proxy tunnel like Cloudflare or ngrok, with password-gated owner approval per client. The pitch is turning ChatGPT int...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.